Modelling Dispersion Using Finite Mixture Of PoissoN
View/ Open
Date
2011Author
Kipchirchir, I. C
Type
ArticleLanguage
enMetadata
Show full item recordAbstract
The negative binomial distribution is a versatile distribution in describing dispersion. The negative binomial parameter k is considered as a dispersion parameter. The aim of this paper is to demonstrate that filnite mixture of Poisson can be used in modelling dispersion. The digamma ." function and the Sterling's expansion for the gamma function are used to construct a dispersion parameter in relation to the negative binomial parameter k. The construction is based on the hierarchical maximum
likelihood estimation of the negative binomial parameter k. The method of moments estimates of
Citation
Kipchirchir I. C. Modelling Dispersion Using Finite Mixture Of Poisson. Far East Journal of Mathematical Sciences. (FJMS) Volume 56, Number 2,2011, Pages 161-178Publisher
University of Nairobi School of Mathematics University of Nairobi